// portfolio
Work & Projects
A collection of projects spanning machine learning, computer vision, full-stack development, and quantum computing.
AutoML — End-to-End Autonomous ML Pipeline
Complete ML engine (profiling → preprocessing → training → tuning → ensembling → explainability → tracking) wrapped in a FastAPI backend and a Streamlit frontend, shipped as a single Docker container.
The Silver House — Full-Stack Jewellery Gallery
Custom e-commerce catalogue for a 925 sterling silver jewellery brand from Srinagar. Built with Next.js 15 (App Router) and Sanity CMS as the backend.
ComplexionAI — AI-Powered Personalized Skincare Intelligence Platform
A Hackathon Winner. ComplexionAI, is a MVP of a full-stack AI-powered app that delivers personalised skincare insights & recommendations based on users' concerns
StyleForge AI — Neural Style Transfer
StyleForge uses Adaptive Instance Normalization (AdaIN) — a feed-forward approach that transfers style in a single forward pass. Decoder network was self-trained entirely, using own training pipeline
SnapClass — AI-Powered Attendance System
SnapClass is a hands-on learning project exploring the intersection of AI, computer vision, and real-world web apps. Not a commercial product — just a developer having fun solving a real problem.
Text Summariser — Fine-tuned T5 on SAMSum
This project fine-tunes a T5-small model to summarize conversations/dialogues into concise summaries. The model is trained on 4,000 dialogue-summary pairs from the SAMSum corpus.
LoanIQ — Machine Learning-Based Loan Eligibility Prediction Pipeline
LoanIQ is a binary classification system that automates loan approval decisions for financial institutions. Data-driven ML pipeline approach for quick customer reviews cutting manual work.
Mini ClaudeCode — Function-Calling AI Agent for Local Codebase Analysis & Engineering
An autonomous Python agent powered by Google Gemini that can explore, analyse, read, write, and execute code in a local codebase - using function calling.
